Collaborates on a team of data scientists to create a more affordable AI based approach to detecting cancerous mutations from genetic data that outperforms current clinical solutions.Successfully implemented a variety of custom AI models to solve various unique tasks using genetic data.Designed and built a machine learning toolkit for the data science team that allows us to quickly and efficiently implement new models to test working hypotheses. This toolkit allows us to share resources across machine learning models, such as metric functions and logging utilities, while utilizing a variety of machine learning frameworks including PyTorch and Keras.Designed and built a dataset toolkit to support our machine learning endeavors along with a clever storage solution on AWS S3 that uniquely fit the team's product requirements.Created new approaches for simulating genetic (Hi-C) samples to improve our models specific deficiencies. Simulations improved performance using a 1D CNN by up to 0.41 mean Average Precision (mAP).

Implemented, trained, and evaluated deep learning algorithms for generating cytogenetics reports from Hi-C heatmaps

Utilized machine learning and gene expression data to model disease states for both the simian immunodeficiency virus and tuberculosis

• Investigated the roles of Sonic hedgehog in Zebrafish skeletal patterning• Formulated and tested hypothesis regarding the role of cell adhesion during skeletal patterning in the zebrafish caudal fin• Developed a novel live imaging technique to investigate cell migration over time• Identified various cell types and the localization of several key molecular factors• Determined specific signaling locations using a photoconvertible protein and live imaging techniques• Designed and assembled a fusion protein gene construct to investigate the exact subcellular localization of a protein• Proposed models explaining the molecular and cellular interactions required for skeletal patterning